Cloud and AI Security Engineer

From Zero to Hired in AI and Cloud Security.
Build it, Break it, Harden it.

Because AI security is cloud security in disguise — you'll master both, and be hireable for either.

Start building — from $37 See what you build
AWS · Terraform · Python Bedrock · MCP · Agents MITRE ATT&CK & ATLAS Portfolio that gets interviews

Two decades in IT. Six in cybersecurity. Taught by a working Cloud & AI Security Architect — not a course factory.

The opportunity

Cloud and AI Security Are Taught Apart. The Best Roles Want Them Together.

Cloud security is established, widely taught, and hiring at volume. AI security is the new frontier, growing fast. Almost every course covers one or the other — so the engineers who can move fluently between the cloud account and the AI running inside it are still rare. That gap is your opening.

It's an opening worth taking, because the two layers are really one system. A model is only as secure as the cloud it runs in — the identity it acts as, the network that exposes it, the key that protects its data, the trail that records what it did. This course teaches both layers as a single skill: you build the cloud foundation first, then build the AI on top of the foundation you own, and you learn to secure the whole thing rather than half of it. Two skills, one engineer, one system.

The method

Build it. Release it. Break it. Harden it.

Nothing in this program is set up secure on the first pass, because nobody learns anything from a finished fortress.

01 · Build

Stand up a deliberately weak version

Deploy it to a real, isolated cloud account. Weak by policy, never broken by omission — it runs, and it's wide open, and both are true at once.

02 · Break

Attack your own build

Run the scanner, run the attacks the threats describe, and document exactly what gets through. Then read the logs and see what never got recorded at all.

03 · Harden

Close every finding

Redeploy hardened, re-run the same scan or attacks, and prove the fix. Read the logs again and watch the same actions light up in the trail.

04 · Prove

Publish the before and after

Tear it down clean, then publish the whole loop as a portfolio artifact. The before-and-after is proof that does the job a resume cannot.

And it can't run up your bill. Every attack that costs money runs against a free mock model, every build ships an always-on spend breaker in code, and you tear every lab down at the end of the session. Labs cost pennies, not dollars — and you learn the professional lesson hiding inside: cloud budgets only alert, they don't hard-stop, so the guardrail lives in code you control.

What you build

A Cloud Security Foundation, Four AI Pillars on Top of It, and the Portfolio That Sells All of It.

You start at the ground and build up — the whole stack, run through the loop on every layer. Nothing floats.

00
The cloud foundationPrivate networking, a customer-managed KMS key, least-privilege IAM, locked-down S3, Secrets Manager, CloudTrail, CloudWatch — built from code in Terraform, deployed weak, scanned, hardened, proven, torn down.
01
Pillar 1 — LLM SecurityA RAG support bot deployed vulnerable, attacked the four ways hiring managers test first, then rebuilt with programmable guardrails and the same attacks proven to fail.
02
Pillar 2 — AI APIs & MCPAn inference API and a naive MCP server in your private subnets: deployed insecure, attacked, then hardened with audience-bound tokens, schema validation, per-token cost caps, and audit logging.
03
Pillar 3 — Agentic AIAn over-privileged agent broken through its own tools and memory, then rebuilt with least-privilege tool scopes, human approval gates, memory integrity, and a kill switch.
04
Pillar 4 — Vibe CodingA codebase full of the holes AI writes by default, proven by scan, then a pipeline rebuilt to catch insecure generation, hardcoded secrets, and the blind spots AI introduces.
05
Threat modeling & detectionMITRE ATT&CK for the attacker's playbook, STRIDE for your own designs, ATLAS for the AI layer, detections mapped to real telemetry, the Capital One breach walked end to end.
06
Your portfolioEvery build, attack, and defense packaged as evidence — with architecture diagrams — plus the job search across both markets: roles, outreach, and interview prep.
The payoff

One Course. Two Job Markets. A Tiny Candidate Pool That Covers Both.

The AI security market is new, high-paying, and small. The cloud security market is established, high-paying, and large. Your training covers both, so your search does too — and while the AI market matures, cloud roles are hiring at volume right now.

AI Security Engineer
$145K–$235K

Total comp, US. Senior and top-of-market roles run well past $350K.

Cloud Security Engineer
$134K–$214K

Typical range, US, ~$169K average — with far more open seats at any moment.

Compensation figures are market estimates (Glassdoor, 2026) for the roles this course targets, not a promise of any specific outcome. The hybrid listings — AI Security roles that require cloud depth, Cloud Security roles that now name LLM and agent security — are your best target, and the pool that covers both sides is small. You'll be in it.

Two ways in

Build It With Me in the Room, or Take the Path on Your Own.

Same curriculum, same method. Full is the whole thing built with me, twenty-two calls across eleven weeks, live and all recorded. Light is the build book on its own, and it upgrades to Full later for the same total.

$399 from 5 November
Full
$136 one time

The whole thing built with me. Everything in Light, plus the calls, the repository and the Python course.

  • Everything in Light, plus:
  • Twenty-two build calls across eleven weeks, two a week. We deploy each pillar weak, attack it, and harden it together
  • Join every call live, and keep all twenty-two recordings whether you make them or not
  • The course repository: Terraform modules, scanners, remediation scripts, runbooks
  • The portfolio week, turning eight builds into the repo a hiring manager opens before they interview you
  • The Python for Security course, a $97 value, included free
  • The Full Discord
Get Full — $136

Instant access. Full goes to $399 on 5 November.

Light Edition
$37 one time

The build book on its own, yours to work at your own pace.

  • The full build book: the foundation and all four AI pillars, every control explained ground-up
  • Deploy-it-weak, scan, harden, prove — the whole loop, printed to build with your own hands
  • Threat modeling, detection, portfolio, and job-search chapters
  • The Discord, where I answer questions myself
  • The course repository
  • The twenty-two build calls, live or recorded
  • The portfolio week
  • The Python for Security course
Get the Light Edition — $37

Upgrade to Full any time for $99. That is $136 either way, so starting here costs you nothing extra.

Every call is recorded and yours to keep, so a week you cannot make live is not a week you lose.

Fit check

Who This Is For.

This is for you if

  • You're moving into cloud or AI security and want a hands-on on-ramp, not theory
  • You're in IT, sysadmin, SOC, AppSec, or dev and want to pivot with real, published work
  • You can read code but want to prove you can build, break, and secure real systems
  • You want a portfolio that changes the interview from "tell me about yourself" to "walk me through your build"

This is not for you if

  • You want a certificate without doing the work
  • You want a passive video course to watch at 2x and forget
  • You're already shipping cloud and AI security systems daily
  • You're not willing to deploy, attack, and harden a single thing yourself
Who built this
Zach Marcy

From Someone Who Does the Work.

Built by Zach Marcy — a Cybersecurity, Cloud & AI Architect, Consultant, and Mentor with 20+ years in IT and 6 in cybersecurity. His day-to-day is designing and securing the cloud environments that deploy and secure APIs and AI, and advising the teams that run them.

The teaching method here is the same one he uses on real systems: build it small, prove it insecure, harden it, and tear it down clean. Every framework in the course comes from work he has actually done. Quality over quantity, depth over noise, real-world relevance above all.

Cybersecurity Education That Gets You Hired, Promoted and Paid.

Questions

Straight Answers.

Do I need cloud or AWS experience to start?

No. The course sets up your AWS sandbox from scratch and explains every service ground-up — IAM, VPC, S3, KMS, CloudTrail — before it's ever asked to carry a lesson. If service names are new to you, that's expected, and it does not put you behind.

Will the AWS labs cost me money?

Designed to cost pennies. Everything runs in an isolated sandbox account, the attacks that cost money run against a free mock model, every build carries an always-on spend breaker in code, and you tear every lab down at the end of the session. You set a budget alarm — but you'll also learn why a budget alarm alerts and doesn't hard-stop, and where the real protection lives.

What's the difference between Light and Full?

Same curriculum and method. Light ($37) is the complete build book you work through on your own, and you can start today. Full ($136) adds the twenty-two build calls live, the course repository, the portfolio week and the Python for Security course, a $97 value, free. You can start with Light and upgrade for $99 later, which is the same $136 total.

Are the calls live, and what if I miss one?

They are live, two a week, Tuesday and Thursday, across eleven weeks. Every call is recorded and yours to keep, so a week you cannot make live is not a week you lose.

What actually happens on the calls?

Two calls a week for eleven weeks. Weeks one and two are the landscape, the jobs, and how to think like an architect. Weeks three to ten are eight builds, and every one runs the same loop: on the first call you deploy it weak and prove it insecure with a scanner, on the second you harden it and prove the same attack now fails. Week eleven turns the eight builds into a portfolio and works the job search. Every call is recorded and yours to keep.

What frameworks and tools does it cover?

Terraform, AWS (IAM, VPC, KMS, S3, CloudTrail, CloudWatch, Bedrock, Secrets Manager, GuardDuty), Python for the scanners and attacks, MITRE ATT&CK and ATLAS, STRIDE, NeMo Guardrails, MCP, and the OWASP LLM Top 10 — mapped to the projects you build.

How does this get me hired?

You walk into interviews with evidence, not aspirations: a hardened foundation and four pillar projects, each with before-and-after scanner output, red-team reports, architecture diagrams, and written threat models — plus a job-search plan across both the cloud and AI security markets.

What's the guarantee?

Work through the first section. If you haven't deployed, scanned, and hardened a real cloud build with your own hands, email me and I'll refund you. No forms, no friction.

Own the ground and the buildings

Build the Whole System —
the Cloud, and the AI on Top of It.

The engineer who can secure both is the one companies trust with their infrastructure, their models, and their integrations — and pay accordingly. This is your path to becoming that engineer, one proven build at a time.

Get Full — $136 Start with Light — $37

Build it. Release it. Break it. Harden it.